Smooth Model Compression without Fine-Tuning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Runkel, Christina, Meli, Natacha Kuete, Lukasik, Jovita, Biguri, Ander, Schönlieb, Carola-Bibiane, Moeller, Michael
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918039894622208
author Runkel, Christina
Meli, Natacha Kuete
Lukasik, Jovita
Biguri, Ander
Schönlieb, Carola-Bibiane
Moeller, Michael
author_facet Runkel, Christina
Meli, Natacha Kuete
Lukasik, Jovita
Biguri, Ander
Schönlieb, Carola-Bibiane
Moeller, Michael
contents Compressing and pruning large machine learning models has become a critical step towards their deployment in real-world applications. Standard pruning and compression techniques are typically designed without taking the structure of the network's weights into account, limiting their effectiveness. We explore the impact of smooth regularization on neural network training and model compression. By applying nuclear norm, first- and second-order derivative penalties of the weights during training, we encourage structured smoothness while preserving predictive performance on par with non-smooth models. We find that standard pruning methods often perform better when applied to these smooth models. Building on this observation, we apply a Singular-Value-Decomposition-based compression method that exploits the underlying smooth structure and approximates the model's weight tensors by smaller low-rank tensors. Our approach enables state-of-the-art compression without any fine-tuning - reaching up to $91\%$ accuracy on a smooth ResNet-18 on CIFAR-10 with $70\%$ fewer parameters.
format Preprint
id arxiv_https___arxiv_org_abs_2505_24469
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Smooth Model Compression without Fine-Tuning
Runkel, Christina
Meli, Natacha Kuete
Lukasik, Jovita
Biguri, Ander
Schönlieb, Carola-Bibiane
Moeller, Michael
Machine Learning
Compressing and pruning large machine learning models has become a critical step towards their deployment in real-world applications. Standard pruning and compression techniques are typically designed without taking the structure of the network's weights into account, limiting their effectiveness. We explore the impact of smooth regularization on neural network training and model compression. By applying nuclear norm, first- and second-order derivative penalties of the weights during training, we encourage structured smoothness while preserving predictive performance on par with non-smooth models. We find that standard pruning methods often perform better when applied to these smooth models. Building on this observation, we apply a Singular-Value-Decomposition-based compression method that exploits the underlying smooth structure and approximates the model's weight tensors by smaller low-rank tensors. Our approach enables state-of-the-art compression without any fine-tuning - reaching up to $91\%$ accuracy on a smooth ResNet-18 on CIFAR-10 with $70\%$ fewer parameters.
title Smooth Model Compression without Fine-Tuning
topic Machine Learning
url https://arxiv.org/abs/2505.24469